Fix pandas FutureWarning and enhance wranglers performance

[rsmb7z]

Pull Request

• Fixed pandas FutureWarning for TradeTickDataWrangler.process_bar_data
• Implemented prepare_event_and_init_timestamps which prepares ts_event and ts_init directly from DatetimeIndex which is already ns. This has improved the performance significantly, less than half time.
• Optional sort_data flag for process_bar_data with default to True

Following is the results on my dataset:

• TradeTickDataWrangler 1970563 rows, processed in 7s (previously 32s)
• QuoteTickDataWrangler 6822199 rows, processed in 13s (previously 112s)

[cjdsellers]

Hey @rsmb7z

The changes look good - I see the PR is in draft mode, did you still intend to add more changes?

[rsmb7z]

Hey @rsmb7z

The changes look good - I see the PR is in draft mode, did you still intend to add more changes?

Hi @cjdsellers
Yes figured out further performance improvements. Will change to 'Ready' once pushed.

[rsmb7z]

Hi @cjdsellers
The performance have further improved as follows.

• 1970563 bars loaded as TradeTick in 7.00s (from 32s)
• 1970563 bars loaded as QuoteTick in 12.89s (from 112s)

I have also added optional flag sort_data set default to True for keeping current behavior. This helps further improve because DataEngine would be sorting eventually as well.